当前位置：首页 > news >正文

python基础入门：8.1项目1：爬虫与数据分析

news 2025/6/20 14:47:14

Python爬虫与数据分析全流程实战：从数据采集到可视化呈现

# 综合案例：电商价格监控分析系统
import requests
from bs4 import BeautifulSoup
import pandas as pd
import matplotlib.pyplot as plt# 配置参数
HEADERS = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) ''AppleWebKit/537.36 (KHTML, like Gecko) ''Chrome/91.0.4472.124 Safari/537.36'
}def scrape_products(url):"""爬取商品信息"""products = []try:response = requests.get(url, headers=HEADERS, timeout=10)soup = BeautifulSoup(response.content, 'html.parser')items = soup.select('div.product-item')for item in items:name = item.select_one('h2.product-title').text.strip()price = item.select_one('span.price').text.strip()rating = item.select_one('div.rating').attrs['data-score']reviews = item.select_one('a.reviews-count').text.split()[0]products.append({'name': name,'price': price,'rating': float(rating),'reviews': int(reviews.replace(',', ''))})except Exception as e:print(f"爬取失败: {str(e)}")return productsdef clean_data(df):"""数据清洗处理"""# 价格处理df['price'] = df['price'].str.replace('$', '').astype(float)# 过滤异常值df = df[(df['price'] > 0) & (df['price'] < 10000)]# 分类处理df['category'] = df['name'].str.extract(r'([A-Za-z]+) Pro')df['category'] = df['category'].fillna('Other')return dfdef visualize_data(df):"""数据可视化展示"""plt.figure(figsize=(15, 8))# 价格分布直方图plt.subplot(2, 2, 1)df['price'].plot(kind='hist', bins=20, color='skyblue')plt.title('价格分布')plt.xlabel('价格 ($)')# 评分与价格散点图plt.subplot(2, 2, 2)plt.scatter(df['rating'], df['price'], alpha=0.6)plt.title('评分 vs 价格')plt.xlabel('评分')plt.ylabel('价格 ($)')# 类别销量柱状图plt.subplot(2, 2, 3)df['category'].value_counts().plot(kind='bar', color='salmon')plt.title('商品类别分布')plt.xticks(rotation=45)# 价格趋势折线图plt.subplot(2, 2, 4)df.sort_values('rating').groupby('rating')['price'].mean().plot(marker='o', color='green')plt.title('不同评分的平均价格')plt.xlabel('评分')plt.ylabel('平均价格 ($)')plt.tight_layout()plt.savefig('product_analysis.png', dpi=300)plt.show()# 主程序
if __name__ == "__main__":# 示例电商网站（需替换实际目标网站）base_url = "https://example-store.com/products?page="all_products = []for page in range(1, 6):  # 爬取前5页url = f"{base_url}{page}"print(f"正在爬取: {url}")all_products.extend(scrape_products(url))df = pd.DataFrame(all_products)df = clean_data(df)print("\n数据概览:")print(df.describe())print("\n保存数据到products.csv")df.to_csv('products.csv', index=False)visualize_data(df)

一、高效爬虫开发技巧

网页解析优化策略

# 使用CSS选择器最佳实践
def optimized_parser(html):soup = BeautifulSoup(html, 'lxml')  # 使用更快的解析器# 选择器优化技巧products = soup.select('div[data-product-id]')  # 通过属性选择for product in products:# 链式查找减少查询次数name = product.find(class_='title').get_text(strip=True)# 使用data属性获取信息price = product.find('meta', {'itemprop': 'price'})['content']# 异常处理try:rating = product.select_one('.stars').attrs['title']except (AttributeError, KeyError):rating = None

反爬虫应对方案

# 高级请求配置
session = requests.Session()
session.proxies = {'http': 'http://10.10.1.10:3128','https': 'http://10.10.1.10:1080',
}# 随机延迟
from random import uniform
from time import sleepdef safe_request(url):sleep(uniform(1, 3))  # 随机延迟1-3秒return session.get(url)# 使用代理中间件示例
class ProxyMiddleware:def process_request(self, request, spider):request.meta['proxy'] = "http://user:pass@proxy_ip:port"

二、数据清洗实战技巧

常见数据问题处理

def advanced_cleaning(df):# 处理缺失值df['rating'] = df['rating'].fillna(df['rating'].median())# 处理重复值df = df.drop_duplicates(subset=['name'], keep='last')# 处理异常值q_low = df['price'].quantile(0.01)q_high = df['price'].quantile(0.99)df = df[(df['price'] > q_low) & (df['price'] < q_high)]# 日期处理df['release_date'] = pd.to_datetime(df['release_date'], errors='coerce', format='%Y-%m')# 文本清洗df['name'] = df['name'].str.replace(r'[^\w\s]', '', regex=True)return df

数据转换技巧

# 创建价格分段
bins = [0, 50, 100, 200, 500, 1000]
labels = ['<50', '50-100', '100-200', '200-500', '500+']
df['price_range'] = pd.cut(df['price'], bins=bins, labels=labels)# 计算价格指数
df['price_index'] = (df['price'] / df.groupby('category')['price'].transform('mean')).round(2)# 时间序列转换
monthly_sales = df.resample('M', on='date')['price'].sum()

三、可视化进阶技巧

交互式可视化（使用Plotly）

import plotly.express as px# 创建交互式散点图
fig = px.scatter(df, x='rating', y='price', color='category',hover_data=['name'], title='商品分布分析')
fig.show()# 创建桑基图（Sankey Diagram）
category_flow = df.groupby(['category', 'price_range']).size().reset_index(name='count')
fig = px.sankey(category_flow, nodes={'label': list(df['category'].unique()) + labels},link=dict(source=category_flow['category'],target=category_flow['price_range'],value=category_flow['count']))
fig.show()

自动化报告生成

from pandas_profiling import ProfileReport# 生成数据分析报告
profile = ProfileReport(df, title="商品数据分析报告")
profile.to_file("product_report.html")# 使用Jupyter Notebook集成
from IPython.display import HTML
HTML(profile.to_html())